Calculadora-BMC

Turn a repeated agent or prod failure into a permanent harness improvement (test/hook/script + optional AGENTS line + RULE-PROVENANCE). Use when user says harness-ratchet, ratchet this failure, never again, or after a live-fix that should become structural.

Complete Shopify store connector for Calculadora BMC: OAuth/PKCE, Admin GraphQL, Storefront API, webhooks, catalog sync, products/variants/inventory/prices, orders, customers, discounts, metafields, theme/Liquid/design, Checkout Extensibility, Shopify Functions, app embeds, and Partner Dashboard options. Use when the user wants to develop Shopify design, functionality, catalog, storefront, theme, app extensions, or any Shopify option from this repo; also for shopify connector, conectar Shopify, tienda Shopify, theme Liquid, Online Store 2.0, or full-store Shopify integration (broader than shopify-integration-v4 questions/quotes).

Implements and maintains the Shopify questions/quotes flow that replaces Mercado Libre: OAuth 2.0/PKCE, webhooks→Google Sheet sync, admin question list, LLM/rule-based suggested replies, one-click approve & send, auto-reply scheduler (UTC-3). Use when working on Shopify backend, Cloud Run, Firestore, Sheets API, HMAC validation, or LATAM Spanish reply formatting for Uruguay.

Orchestrates local Agent Review Issue and Fix: Bugbot review (optional), bmc-issue-fix-reviewer fixes, optional security-review, expert-debug cleanup, and gate:local. Use when user says /issue-and-fix, issue and fix, revisar y arreglar, agent review local, or Cursor Agent Review insufficient funds.

Local alternative to Cursor Agent Review (Issue and Fix): reviews branch or uncommitted diffs against project rules, finds bugs/security/logic issues, and applies fixes. Runs gate:local after fixes. Use when Agent Review fails (insufficient funds), user says issue and fix, revisar y arreglar, or review and fix changes.

Remotion video creation expertise — compositions, useCurrentFrame, interpolate, Sequences, assets via staticFile, render CLI, captions, audio, FFmpeg. Use when working with .tsx files importing from "remotion" or "@remotion/*", building video compositions in React, configuring Root.tsx, rendering frames/stills, adding audio/video/captions, or when the user mentions Remotion, Remotion Studio, or video composition. Do NOT use for general React UI work, non-video animation, or pure ffmpeg pipelines without React.

Run promptfoo evals against BMC chat prompts (Panelin agent, autolearn, ML responses). Use when the user wants to test, compare, or regression-check LLM prompts/models in this repo.

Apply the Anthropic "Applied AI" frontend visual language to BMC components — calm warm palette, generous whitespace, serif display + clean sans body, restrained color, content-first layout. Use when redesigning a screen, creating a new panel, or auditing existing JSX/CSS for visual consistency. Trigger when user mentions "Applied AI design", "Anthropic style", "claude.ai look", or asks to redesign/restyle a BMC component.

Evidence-based holistic snapshot of Calculadora BMC: project status from PROJECT-STATE, architecture areas (Vite SPA, Express API, Sheets/MATRIZ, ML/Shopify, Drive, Postgres, hub UI, GPT/MCP, CI), health commands (smoke:prod, gate:local, contracts, ml:verify), readiness % per area with explicit legend, recent developments, risks, and next steps. Use when the user asks for project health, architecture map, readiness report, full status snapshot, or interpret smoke/gate results across modules.

Positions the prompt-receiving agent as the orchestrator for a task: designs per-terminal work, issues expert instruction prompts per track, polls log files until gates match, cross-ingests findings across tracks, and live-rebalances instructions so all tracks stay aligned. Uses coprocesses, .runtime/orch logs, gates, and HIL. Use when the user wants a central orchestrator, multi-terminal coordination, expert prompts per pane, cross-track discovery merge, or equitable shared objective across parallel work.

Orchestrates multi-terminal and multi-session work: assigns roles, env, and run order; waits on logs, ports, and exit codes; reconciles live terminal output. Also synthesizes chaos from several Cursor terminals or agent sessions. Use when the user wants parallel panes, process orchestration, live logs, or to merge many terminal attachments into one runbook.

BMC Team Liaison: helps humans and agents align on intent, repo conventions, and the right specialist. Refines vague asks into a crisp brief; points to AGENTS.md, human gates, Contribut mode, PROJECT-STATE, and PROJECT-TEAM-FULL-COVERAGE. Suggests which Cursor agent or skill to invoke next. Use when the user asks for a copilot, liaison, alignment help, "who should do this", or communication help across BMC/Panelin work.

Ejecuta tareas bajo un runbook acordado y, al cerrar, sintetiza procedimiento, decisiones, orden y excepciones para persistir como conocimiento reutilizable del asistente (log/KB/skill). Usar cuando el usuario activa RUNBOOK, modo procedimiento aprendizaje, capturar lógica operativa, cerrar corrida con aprendizajes, o enriquecer conocimiento del bot tras ejecutar pasos repetibles.

Two-phase chat workflow: refine and enrich the user's raw message into a clear "official input" before doing heavy work. Activate with CONTRIBUT ON or phrases like "modo Contribut", "Contribut", "refinar mi input antes de responder". Deactivate with CONTRIBUT OFF. One-shot bypass: CONTRIBUT SKIP. Gate to full reply: user sends ACEPTO BORRADOR (optionally pasting the refined text). Use when the user wants higher input precision and alignment before execution.

Implementación y verificación end-to-end de Google Drive (GIS) en Calculadora-BMC: OAuth Web client, VITE_GOOGLE_CLIENT_ID, scripts drive:bootstrap / drive:configure / drive:vercel-env, verify de env y de dist, workflows GitHub opt-in. Usar cuando el usuario pida Drive, invalid_client, VITE_GOOGLE_CLIENT_ID, automatizar setup GCP/Vercel, o “agente Drive OAuth”.

Extrae documentación y arquitectura desde Sketchfab Developers (OAuth2, Data/Download/Viewer API, oEmbed) y la línea comercial 3D Configurators (aplicaciones custom sobre Viewer API; no código fuente del configurador listo para copiar). Incluye Quantifier Pro (plugin SketchUp) y SketchUp Community Forums (evidencia con nivel de confianza). Usar cuando el usuario pida reverse engineering de documentación, mapa de arquitectura, integración Sketchfab, 3D configurators, Viewer API, OAuth Sketchfab, Quantifier Pro, takeoff o costos en SketchUp, forums.sketchup.com, o distinguir API oficial vs foro.

Conector Mercado Libre en Calculadora-BMC: arranque de la API (puerto 3001), OAuth 2.0 (UY), ngrok/HTTPS, variables .env, verificación (ml:verify), y uso de rutas /auth/ml/* y /ml/* (preguntas, órdenes, users/me). Usar cuando el usuario o el flujo requieran Mercado Libre, Panelin ML, OAuth ML, ngrok, curl a localhost:3001/ml, o cambios en server/index.js / mercadoLibreClient.

Ciclo operativo Mercado Libre (BMC): levantar API y OAuth, exportar corpus de preguntas, priorizar UNANSWERED, iterar la KB de respuestas con el humano, cotizar con /calc/cotizar, normalizar moneda (U$S→USD) y publicar con POST /ml/questions/:id/answer. Usar cuando pidan operar preguntas ML, mejorar ML-RESPUESTAS-KB-BMC, publicar respuestas al vendedor, o repetir el flujo “servicios listos → KB → ML” sin improvisar pasos.

ISO 128/129/3098 drawing quality skill for the Panelin 2D roof plan, 3D realistic preview, and print/PDF pipeline. Provides a 100-point scoring rubric across 7 domains, implementation patterns, and verification workflow.

Orchestrates propagation, PROJECT-STATE updates, and local vs production verification for Calculadora BMC (https://calculadora-bmc.vercel.app): Vite localhost:5173, API localhost:3001, npm run smoke:prod against public API, contracts/capabilities when routes or OpenAPI change. Use when the user asks to sync the calculadora with prod, align local and remote, propagate changes across API/integrations/MATRIZ, run cross-area sync for Panelin calc, or verify Vercel plus Cloud Run after edits.

CEO AI Agent — leads the BMC/Panelin project with a single objective: project working by end of week. Invokes full team run repeatedly until the goal is achieved. Use when user says CEO run, CEO agent, project working by end of week, or make it work.

Enables interactive chat/dialogue with the full BMC team. User asks questions, agent coordinates responses from the right members, allows follow-ups, and can invoke full team when needed. Use when user wants chat con el equipo, diálogo interactivo, or conversational team coordination.

Configure and integrate Meta platform APIs (WhatsApp Business Platform, Instagram Graph API, and Facebook Graph API) end-to-end for an application, including app setup, auth tokens, webhooks, permissions, backend endpoints, testing, and production hardening. Use when asked to build or debug social messaging/posting integrations, webhook flows, or Meta developer console configuration.

Panelin-Gym: preparar el entorno local y el flujo repetible para observabilidad del asistente (chat calculadora + canal ML), KB en disco, edición de prompts vía API/UI dev, import masivo, corpus/export ML y buenas prácticas de entrenamiento operativo (sin fine-tuning de pesos). Usar cuando el usuario diga Panelin-Gym, gym Panelin, entrenar Panelin, observabilidad del chat, revisar conversaciones/KB, corpus ML, gold runs, o alinear prompts con datos reales.

Actualiza los precios de la Calculadora BMC desde la MATRIZ de COSTOS y VENTAS 2026. Lee BROMYROS: F/L/T tal cual; M ref. c/IVA tal cual; **col U** = venta web **c/IVA** → CSV `venta_web_iva_inc` (tal cual; sin push); genera CSV (path, costo, venta_local, venta_web, …). Use cuando el usuario pida: actualizar precios calculadora, sincronizar precios desde MATRIZ, levantar precios de la planilla, /actualizar-precios-calculadora, o consultar precios desde la matriz.

Convierte transcripción de navegación (voz/texto) en beats U-xx, cruza con evidencia MCP chrome-devtools (E-xx), y produce plan de acción + parches; repetible para Calculadora BMC local o prod.

Turns Panelin Signal / Knowledge Antenna outputs into evaluated improvement quests and concrete implementation plans for Calculadora BMC (https://calculadora-bmc.vercel.app). Reads magazine HTML and underlying JSON reports, maps news to repo areas, and produces scoped plans (full stack or single area). Use when the user asks for improvements from knowledge news, Panelin Signal, antenna magazine, knowledge run output, or implementation plans driven by external AI/platform updates.

Live DevTools narrative (Narrativa en vivo DevTools): sesión en Cursor donde el agente usa el MCP chrome-devtools para navegar la app (default calculadora-bmc.vercel.app), extrae consola/red/performance/DOM según aplique, y cruza en tiempo real con la narrativa del usuario (chat o transcripción pegada), organizada para lectura. Use when the user says Live DevTools narrative, Narrativa en vivo DevTools, or wants MCP browser evidence linked to their commentary on the Calculadora BMC Vercel app.

Video-User-interactive-dev: método nombrado por el usuario para analizar un vídeo de sesión (path o adjunto), revisión completa documentada y procedimiento de desarrollo en un solo informe. También: sesiones MP4/MOV (iPhone, Loom, OBS), metadatos, JSON + Markdown en docs/team/ux-feedback. Use when the user says Video-User-interactive-dev, shares a session video path, or wants USER POV video analysis and a dev plan in Calculadora-BMC.

Ante disco lleno (ENOSPC, no space left on device), audita espacio en modo seguro, propone liberación por GRUPOS (origen, impacto, fechas típicas), pide aprobación explícita del usuario, ejecuta lo aprobado y reanuda el trabajo que quedó bloqueado. Use when disk is full, write fails for lack of space, or the user asks to free space and continue a blocked procedure.

Convierte transcripción de audio, capturas y URL de la app desplegada en un informe Markdown estructurado (USER-NAV-REPORT) con hallazgos, propagación multi-área y backlog accionable para un agente implementador. Use when the user shares voice transcript, screenshots, or navigation feedback about the Vercel/Panelin app and wants a structured report incorporable into the project.

Ejecuta desde Cursor (agente con terminal) verificación local, smoke a producción, checks pre-deploy y guía de deploy Cloud Run (panelin-calc) + Vercel para Calculadora BMC. Usar cuando pidan deploy desde este entorno, dejar configurado el flujo, redeploy API/MATRIZ, o “hacerlo todo desde Cursor”.

Revisa la tarea y la intención del usuario (incluyendo matices entre paréntesis), analiza prompts y brechas, escanea soluciones open source, extensiones y bases reutilizables, y propone mejoras orientadas a las mejores capacidades de producto. Use when refining intent, reviewing a spec before build, comparing OSS options, or choosing foundations to ship faster with better features.

Asesor macOS para rendimiento estable, menos beach balls y diagnóstico ordenado (CPU, memoria, disco, energía, red). Pasos seguros primero; avanzado solo con backup y confirmación. Use when the user asks to speed up Mac, fix hangs, reduce lag, optimize macOS, Activity Monitor, or system freezes.

See a local page in the browser while editing files in Cursor: run the right dev server (Vite HMR or static live-reload), dock Simple Browser beside the editor, and keep URL + port in sync so saves reflect almost instantly. Use when the user wants to watch a local browser tab, live preview, instant refresh while coding, or edit HTML/CSS/JS here and see changes immediately.

Syncs project state across all BMC/Panelin areas. Reads PROJECT-STATE.md, runs propagation when changes affect multiple agents, updates state and pendientes. Use when user says sync project state, actualizar equipo, full team run, or keep everyone updated after a change.

Organiza carpetas y documentación en Calculadora-BMC, detecta huecos o docs obsoletos, mantiene índices en docs/ y READMEs, y guía flujos GitHub (ramas, PR, etiquetas). Use when the user asks to organize documentation, fix doc structure, update READMEs, doc index, repo hygiene, or GitHub workflow for docs.

Evalúa según mejores desempeños en distintas áreas y tareas. Sabe desde cero qué procesos conviene ejecutar en paralelo o en serie. Muy orientado a objetivos; prevé según scores y contexto la mejor combinación de agentes. Use when deciding parallel vs serial execution, agent combination, or orchestration strategy.

Por cada run: estado interno + WATCHLIST, descubrimiento de grupos/canales públicos relacionados, escaneo de contenido permitido (export/Bot API), verificación de enlaces y memo de decisiones de implementación (hacer / piloto / diferir / rechazar). Use when the user wants recurring Telegram intelligence for BMC/Panelin roadmap decisions.

MATPROMT — Optimización de prompts y flujos de trabajo para el equipo BMC. Genera instrucciones específicas por rol para full team run, refina prompts ante tareas nuevas durante el run, y mejora handoffs. Usar cuando se invoque full team, se pida MATPROMT, o se necesiten prompts orientadores por agente.